arxiv: v3 [cs.cg] 20 Jul 2017

Transcription

1 An annotated bibliography on 1-planarity Stephen G. Kobourov 1, Giuseppe Liotta 2, Fabrizio Montecchiani 2 1 Dep. of Computer Science, University of Arizona, Tucson, USA kobourov@cs.arizona.edu 2 Dip. di Ingegneria, Università degli Studi di Perugia, Perugia, Italy {giuseppe.liotta,fabrizio.montecchiani}@unipg.it arxiv: v3 [cs.cg] 20 Jul 2017 July 21, 2017 Abstract The notion of 1-planarity is among the most natural and most studied generalizations of graph planarity. A graph is 1-planar if it has an embedding where each edge is crossed by at most another edge. The study of 1-planar graphs dates back to more than fifty years ago and, recently, it has driven increasing attention in the areas of graph theory, graph algorithms, graph drawing, and computational geometry. This annotated bibliography aims to provide a guiding reference to researchers who want to have an overview of the large body of literature about 1-planar graphs. It reviews the current literature covering various research streams about 1-planarity, such as characterization and recognition, combinatorial properties, and geometric representations. As an additional contribution, we offer a list of open problems on 1-planar graphs. 1 Introduction Relational data sets, containing a set of objects and relations between them, are commonly modeled by graphs, with the objects as the vertices and the relations as the edges. A great deal is known about the structure and properties of special types of graphs, in particular planar graphs. The class of planar graphs is fundamental for both Graph Theory and Graph Algorithms, and is extensively studied. Many structural properties of planar graphs are known and these properties can be used in the development of efficient algorithms for planar graphs, even where the more general problem is NP-hard [11]. Most real world graphs, however, are non-planar. For example, scale-free networks (which can be used to model web-graphs, social networks and biological networks) consist of sparse non-planar graphs. To analyze such real world networks, we need to address fundamental mathematical and algorithmic challenges for sparse non-planar graphs, which we call beyond-planar graphs. Beyond-planar graphs are more formally defined as non-planar graphs with topological constraints, such as forbidden crossing patterns, as is the case with graphs where the number of crossings per edge or the number of mutually crossing edges is bounded by a constant (see, e.g.,[74, 111, 119]). A natural motivation for studying beyond-planar graphs is visualization. The goal of graph visualization is to create a good geometric representation of a given abstract graph, by placing all the vertices and routing all the edges, so that the resulting drawing represents the graph 1

2 well. Good graph drawings are easy to read, understand, and remember; poor drawings can hide important information, and thus may mislead. Experimental research has established that good visualizations have a number of geometric properties, known as aesthetic criteria, such as few edge crossings, symmetry, edges with low curve complexity, and large crossing angles; see [86, 112, 113, 114]. Visualization of large and complex networks is needed in many applications such as biology, social science, and software engineering. A good visualization reveals the hidden structure of a network, highlights patterns and trends, makes it easy to see outliers. Thus a good visualization leads to new insights, findings and predictions. However, visualizing large and complex real-world networks is challenging, because most existing graph drawing algorithms mainly focus on planar graphs, and consequently have made little impact on visualization of real-world complex networks, which are non-planar. Therefore, effective drawing algorithms for beyond-planar graphs are in high demand from industry and other application domains. The notion of 1-planarity is among the most natural and most studied generalizations of planarity. A graph is 1-planar if there exists an embedding in which every edge is crossed at most once. The family of 1-planar graphs was introduced by Ringel [115] in 1965 in the context of the simultaneous vertex-face coloring of planar graphs. Specifically, if we consider a given planar graph and its dual graph and add edges between the primal vertices and the dual vertices, we obtain a 1-planar graph. Ringel was interested in a generalization of the 4-color theorem for planar graphs, he proved that every 1-planar graph has chromatic number at most 7, and conjectured that this bound could be lowered to 6 [115]. Borodin [23] settled the conjecture in the affirmative, proving that the chromatic number of each 1-planar graph is at most 6 (the bound is tight as for example the complete graph K 6 is 1-planar and requires six colors). The aim of this paper is to present an annotated bibliography of papers devoted to the study of combinatorial properties, geometric properties, and algorithms for 1-planar graphs. Similar to [52], this annotated bibliography reports the references in the main body of the sections; we believe that this choice can better guide the reader through the different results, while reading the different sections. All references are also collected at the end of the paper, in order to give the reader an easy access to a complete bibliography. Paper organization The remainder of this paper is structured as follows. Section 2 contains basic terminology, notation, and definitions. It is divided in two main parts as follows. Section 2.1 introduces basic definitions and notation about graphs, drawings, and embeddings; this section can be skipped by readers familiar with graph theoretic concepts and graph drawing. Section 2.2 contains more specific definitions for 1-planar graphs and for subclasses of 1-planar graphs such as IC-planar and NIC-planar graphs. Because different papers on 1-planarity often use different terminology and notation, this section is important for the rest of the paper. Section 3 is concerned with two fundamental and closely related aspects of 1-planar graphs. Specifically, Section 3.1 and Section 3.2 survey known results for the problem of characterizing and recognizing 1-planar graphs, respectively. While the recognition problem is NP-complete for general 1-planar graphs, interesting characterizations and polynomial-time recognition algorithms are known for some meaningful classes of 1-planar graphs, such as optimal 1-planar graphs. Section 4 investigates structural properties of 1-planar graphs, i.e., properties that depend only on the abstract structure of these graphs. Specifically, this section contains: the main results on the chromatic number, the chromatic index, and on other coloring parameters 2

3 (Section 4.1); upper and lower bounds on the edge density for several classes of 1-planar graphs (Section 4.2); results on edge partitions with specific properties on the edge sets (Section 4.3); bounds on various graph parameters (Section 4.4); results on the existence of subgraphs with bounded vertex degree (Section 4.5); results on binary graph operations that preserve 1-planarity (Section 4.6). Section 5 describes several results about geometric representations of 1-planar graphs, which are of interest in computational geometry, graph drawing and network visualization. In particular, the following types of representation are considered: straight-line drawings (Section 5.1); drawings with right-angle crossings (Section 5.2); visibility representations (Section 5.3); and contact representations (Section 5.4). 2 Preliminaries 2.1 Basic definitions In this section we introduce some preliminaries and definitions about graphs and about drawings and embeddings of graphs. For more details about graph theory, the reader is referred to classic textbooks such as: [21] J. A. Bondy and U. S. R. Murty. Graph theory. Graduate texts in mathematics. Springer, [59] R. Diestel. Graph Theory, 4th Edition, volume 173 of Graduate texts in mathematics. Springer, [76] A. Gibbons. Algorithmic Graph Theory. Cambridge University Press, [81] F. Harary, editor. Graph Theory. Addison-Wesley, Further graph drawing background can also be obtained in several books and surveys on the topic, including: [53] G. Di Battista, P. Eades, R. Tamassia, and I. G. Tollis. Graph Drawing: Algorithms for the Visualization of Graphs. Prentice-Hall, [91] M. Jünger and P. Mutzel, editors. Graph Drawing Software. Springer, [94] M. Kaufmann and D. Wagner, editors. Drawing Graphs, Methods and Models, volume 2025 of LNCS. Springer, [109] T. Nishizeki and M. S. Rahman. Planar graph drawing, volume 12 of Lecture notes series on computing. World Scientific, [118] K. Sugiyama. Graph Drawing and Applications for Software and Knowledge Engineers, volume 11 of Series on Software Engineering and Knowledge Engineering. World Scientific, [122] R. Tamassia. Advances in the theory and practice of graph drawing. Theor. Comput. Sci., 217(2): , [123] R. Tamassia, editor. Handbook on Graph Drawing and Visualization. Chapman and Hall/CRC, In order to make this paper self-contained, we review several basic definitions that will be used in the next sections. 3

4 2.1.1 Graphs A graph is a mathematical discrete structure used to represent a set of interconnected objects. Formally, a graph G = (V, E) is an ordered pair comprising a finite and non-empty set V of elements called vertices and a finite set E of elements called edges. An edge e E is an unordered pair (u, v) of vertices; vertices u and v are called end-vertices of e. We often write e = (u, v) to denote an edge e with end-vertices u and v. Vertices u and v are said to be adjacent, while e is incident to all of its end-vertices. The set of vertices adjacent to a vertex v is called the neighborhood of v and denoted by N(v). The number of edges that are incident to a vertex v is called the degree of v and is denoted by deg(v). A graph G has maximum vertex degree, if deg(v) for every vertex v of G and there is a vertex u of G such that deg(u) =. An edge of G is a self-loop if its end-vertices coincide. Also, G contains multiple edges if it has two or more edges with the same end-vertices. A graph G is simple if it has neither self-loops nor multiple edges. Note that in a simple graph the degree of a vertex v corresponds to the size of N(v). In the following we always refer to simple graphs, unless specified otherwise. A subgraph G of a graph G = (V, E) is a graph G = (V, E ), such that V V and E E. If V is a subset of V, the subgraph of G induced by V is the graph G = (V, E ), where E E is the subset of all edges in E connecting any two vertices that are in V. A subgraph G = (V, E ) of G = (V, E) such that V = V is a spanning subgraph of G. A path is a graph whose set of vertices is V = {v 1,..., v n } and there exists an edge e i = (v i, v i+1 ) for i = 1,..., n 1. A component G of a graph G is a maximal subgraph of G such that for every pair u, v of vertices of G there is a path between u and v in G. A graph that has exactly one component is said to be connected. A separating k-set, k 1, of a graph G is a set of k vertices whose removal increases the number of connected components of G. Separating 1-sets and 2-sets are called cut vertices and separation pairs, respectively. A connected graph is 2-connected if it has no cut vertices. A 2-connected graph is 3-connected if it has no separation pairs. In general, a graph is k-connected, k 2, if it has no separating (k 1)-sets. Clearly, if a graph is k-connected, then it is also (k 1)-connected. A connected graph G is a cycle if every vertex of G has degree 2. A graph is acyclic if it does not contain subgraphs that are cycles. A connected acyclic graph is called a tree. A collection of trees (i.e., an acyclic graph) is a forest. A graph is complete if it has an edge connecting every pair of vertices. A complete graph with n vertices is denoted by K n. A graph G is k-partite if the set of its vertices can be partitioned into k 2 sets (also called parts), V 1,..., V k, such that every edge of G connects a vertex in V i to a vertex in V j, for some 1 i j k. A graph G is complete k-partite if it is k-partite and every two vertices that belong to distinct parts are adjacent. A 2-partite graph is also called a bipartite graph. A complete bipartite graph such that V 1 = n 1 and V 2 = n 2 is denoted by K n1,n Drawings and Embeddings Let G = (V, E) be a graph, a drawing on the plane, or simply a drawing, Γ of G is a pair of functions Γ V, Γ E, where Γ V maps each vertex v to a point Γ V (v) of the plane and Γ E maps each edge (u, v) to a simple open Jordan curve Γ E (u, v), whose endpoints are Γ V (u) and Γ V (v). The curves representing the edges are allowed to intersect, but they may not pass through vertices except for their endpoints. In what follows, we will often refer to the vertices and to the edges of a drawing as an abbreviation for the points and the curves that represent the vertices and the edges of the depicted graph. 4

5 A drawing is planar if there are no two edges that cross (except at common endpoints). A planar graph is a graph that admits a planar drawing. A drawing divides the plane into topologically connected regions, called faces. The infinite region is called the outer face. For a planar drawing the boundary of a face consists of vertices and edges, while for a non-planar drawing the boundary of a face may contain vertices, crossings, and edges (or parts of edges). Given a drawing Γ, the planarization Γ p of Γ is the drawing obtained by replacing each crossing point of Γ with a dummy vertex. The vertices of Γ in Γ p are called original vertices to avoid confusion. A drawing where all the edges are mapped to segments is called a straight-line drawing. A polyline drawing is a drawing where the edges are mapped to chains of segments. A drawing, either straight-line or polyline, where all the vertices and all the possible bend points are mapped to points with integer coordinates is a grid drawing. The bounding box of a grid drawing Γ is the minimum axis-aligned box containing Γ. If the bounding box has side lengths X 1 and Y 1, then we say that Γ has area X Y. A rotation system of a graph is defined by an assignment of a clockwise order of the edges incident to a vertex, for all vertices of the graph. A drawing is said to respect a rotation system if scanning around a vertex in clockwise order encounters the edges in the prescribed order. An embedding of a graph G is an equivalence class of drawings of G that define the same set of faces and the same outer face. A planar embedding is an embedding that represents an equivalence class of planar drawings. Equivalently, an embedding of a graph is an equivalence class of drawings under homeomorphisms of the plane. For planar graphs the information carried in an embedding is equivalent to that in a rotation system together with one face indicated to be the outer face. More in general, a graph with a fixed embedding comes with a rotation system, a clockwise order of (part of) edges around each crossing, and with one face indicated to be the outer face. A plane graph is a graph with a fixed planar embedding. Similarly as for a drawing, given a graph G with a fixed embedding, the planarization G p of G is the graph obtained by replacing each crossing of G with a dummy vertex. The vertices of G in G p are called original vertices to avoid confusion. The planarization G p is a plane graph, i.e., a planar graph with a fixed embedding without crossings. It is worth recalling that a 3-connected planar graph has a unique planar embedding up to the choice of the outer face [131]. [131] H. Whitney. Congruent graphs and the connectivity of graphs. Am. J. Math., 54(1): , planar Graphs A 1-planar drawing is one in which each edge is crossed at most once. A graph is 1-planar if it admits a 1-planar drawing. A 1-planar embedding is an embedding that represents an equivalence class of 1-planar drawings. A 1-plane graph is a graph with a fixed 1-planar embedding. A 1-planar graph G with n vertices has at most 4n 8 edges [19, 111] (see also Section 4.2). If G has exactly 4n 8 edges, then it is called optimal. A 1-planar graph G is maximum if it has the maximum number of edges over all 1-planar graphs with n vertices. Note that there exist graphs that are maximum but not optimal (e.g., the complete graph with 5 vertices), whereas an optimal 1-planar graph is also maximum. A 1-planar graph G is maximal if no edge can be added to G without having at least one edge crossed twice in every drawing of G. A 1-planar graph is triangulated if it admits a drawing in which every face contains either exactly three distinct vertices or exactly two distinct vertices and one crossing point. [19] R. Bodendiek, H. Schumacher, and K. Wagner. Bemerkungen zu einem Sechsfarbenproblem von G. Ringel. Abhandlungen aus dem Mathematischen Seminar der Universitaet 5

6 Hamburg, 53(1):41 52, [111] J. Pach and G. Tóth. Graphs drawn with few crossings per edge. Combinatorica, 17(3): , Let Γ be a 1-planar drawing, and let Γ p be its planarization. The drawing Γ is of of class C x, for x {0, 1, 2}, if for every two dummy vertices u 1 and u 2, it holds that N(u 1 ) N(u 2 ) x. A 1-planar graph G is of class C x, for x {0, 1, 2}, if it has a drawing of class C x and no drawing of class C x 1 (if x > 0). A 1-planar graph of class C 0 is also called an IC-planar graph, where IC stands for independent crossings (see, e.g., [5]). A 1-planar graph of class C 1 is also called a NIC-planar graph, where NIC stands for near-independent crossings (see, e.g., [135]). Czap and Šugerek [49] observed that every 1-planar graph belongs to one of the classes C 0, C 1, or C 2. The concepts of IC-planar and NIC-planar embedding, and of IC-plane and NIC-plane graph, can be defined similarly as for the 1-planar case. [5] M. O. Albertson. Chromatic number, independence ratio, and crossing number. Ars Math. Contemp., 1(1), [49] J. Czap and P. Šugerek. Drawing graph joins in the plane with restrictions on crossings. Filomat, 31(2): , [135] X. Zhang. Drawing complete multipartite graphs on the plane with restrictions on crossings. Acta Math. Sin. English Series, 30(12): , A drawing is outer 1-planar if it is 1-planar and all the vertices belong to the outer face of the drawing. An outer 1-planar graph is a graph that admits an outer 1-planar drawing. One can define the concepts of outer 1-planar embedding and outer 1-plane graph similarly as above. 3 Characterization and Recognition In this section we deal with the problems of characterizing and recognizing 1-planar graphs. We recall that finding a characterization for a family F of graphs means finding a property P such that a graph G has the property P if and only if G F, while the recognition problem for a family of graphs F is the problem of deciding whether a graph G belongs to F. The two problems are closely related. In particular, if a property P can be verified in polynomial time on a graph G and P characterizes a family of graphs F, then deciding whether G F can be done in polynomial time. We investigate these two problems for general 1-planar graphs and for several subfamilies of 1-planar graphs. 3.1 Characterization In what follows we survey results concerning the characterization of general 1-planar graphs (Section 3.1.1), optimal 1-planar graphs (Section 3.1.2), triangulated 1-planar graphs (Section 3.1.3), and complete k-partite 1-planar graphs (Section 3.1.4) General 1-planar graphs The class of 1-planar graphs is not closed under edge-contraction, and thus 1-planar graphs are not minor closed [40]. For example, the n n 2 grid graph which is 1-planar and contains K n as a minor, see [63]. 6

7 (a) Q (b) G Figure 1: (a) A 3-connected planar quadrangulation Q and (b) the optimal 1-planar graph G constructed by inserting a pair of crossing edges in each face of Q. [40] Z. Chen and M. Kouno. A linear-time algorithm for 7-coloring 1-plane graphs. Algorithmica, 43(3): , [63] V. Dujmović, D. Eppstein, and D. R. Wood. Structure of Graphs with Locally Restricted Crossings. CoRR, abs/ , To appear in J. Discr. Math. A minimal non-1-planar graph is a graph that is not 1-planar and such that all its proper subgraphs are 1-planar. Korzhik and Mohar [96] proved that the number of n-vertex distinct (i.e., non-isomorphic) minimal non-1-planar graphs is exponential in n (for n 63). These results can be seen as an indication that a characterization of 1-planar graphs by Wagner-type theorem using a finite number of forbidden minors is unlikely. Similarly, for any graph there exists a 1-planar subdivision, which implies the impossibility of having a characterization of 1-planar graphs by Kuratowski-type theorem using a finite number of forbidden subdivisions. [96] V. P. Korzhik and B. Mohar. Minimal obstructions for 1-immersions and hardness of 1-planarity testing. J. Graph Theory, 72(1):30 71, Optimal 1-planar graphs We recall that an n-vertex 1-planar graph is optimal if it contains 4n 8 edges (see also Section 2.2). A planar quadrangulation is a planar graph that admits a drawing in which each face f has four vertices. Let Q be a 3-connected planar quadrangulation. Since Q is 3-connected, it has a unique planar embedding up to the choice of the outer face, and thus a uniquely defined set of faces (or, equivalently, a unique planar embedding on the sphere). Let f be a face of Q, such that in a closed walk along its boundary we encounter its four vertices in this order: u 1, u 2, u 3, u 4. The operation of inserting a pair of crossing edges in f means adding the edges (u 1, u 3 ) and (u 2, u 4 ) inside f such that they cross each other. Note that the two edges do not exist in G (since G is bipartite). Note also that the resulting graph is clearly 1-planar. Bodendiek et al. [19, 20] (and later Suzuki [121]) proved that a 1-planar graph G is optimal if and only if G can be obtained from a 3-connected planar quadrangulation Q by inserting a pair of crossing edges inside each face of Q; see Figure 1 for an illustration. Based on this characterization, it is possible to prove that there are no n-vertex optimal 1-planar graphs with n 7 or n = 9 vertices, while n-vertex optimal 1-planar graphs always exist with n = 8 and n 10 vertices [19, 20]. [19] R. Bodendiek, H. Schumacher, and K. Wagner. Bemerkungen zu einem Sechsfarbenproblem von G. Ringel. Abhandlungen aus dem Mathematischen Seminar der Universitaet Hamburg, 53(1):41 52,

8 u 1 u 2 u 1 u 2 u 1 u 2 u 2 v 1 v 2 u 4 u 3 u 4 v 4 v 3 (a) u 4 u 3 (b) u 4 u 3 u 1 u 2 u 1 u 2 u 1 u 2 u 2 v 1 v 2 u 4 u 3 u 4 v 4 v 3 (c) u 4 u 3 (d) u 4 u 3 (e) W 8 (f) XW 8 Figure 2: The face contraction and vertex splitting operations for (a) planar quadrangulations and (c) optimal 1-planar graphs. The 4-cycle addition and 4-cycle removal operations for (b) planar quadrangulations and (d) optimal 1-planar graphs. (e) The pseudo double wheel W 8. (f) The X-pseudo double wheel XW 8. [20] R. Bodendiek, H. Schumacher, and K. Wagner. Uber 1-optimale Graphen. Mathematische Nachrichten, 117(1): , [121] Y. Suzuki. Optimal 1-planar graphs which triangulate other surfaces. Discrete Math., 310(1):6 11, The above characterization establishes a one-to-one mapping between the set of 3-connected planar quadrangulations and the set of optimal 1-planar graphs. A similar family of 1-planar graphs are the kinggraphs [42]. They are defined as those 1-planar graphs that can be generated by inserting a pair of crossing edges in each quadrangular inner face of a square-graph (i.e., a plane graph in which all inner faces have four vertices and every such vertex has degree at least 4). These 1-planar graphs generalize the graphs that represent all legal moves of the king chess piece on a chessboard, called King s graphs [36]. Let Q be a 3-connected planar quadrangulation. Let f be a face of Q such that in a closed walk along its boundary we encounter its four vertices in this order: u 1, u 2, u 3, u 4. The operation of face contraction of f at {u 1, u 3 } in Q identifies u 1 and u 3 and replaces the two pairs of multiple edges (u 1, u 2 ), (u 2, u 3 ) and (u 1, u 4 ), (u 3, u 4 ) with two single edges, respectively. The inverse operation is called a vertex-splitting; see Figure 2(a) for an illustration. These two operations 8

9 can be applied only if they do not break the simplicity or the triconnectivity of the graph. A 4-cycle addition to f is the operation of adding a 4-cycle v 1, v 2, v 3, v 4 inside f and adding the crossing-free edges (u i, v i ), for i = 1, 2, 3, 4. The inverse operation is called 4-cycle removal; see Figure 2(b) for an illustration. This last operation can be applied only if it does not break the triconnectivity of the graph. A planar quadrangulation is irreducible if neither face contraction nor 4-cycle removal can be applied to H. The pseudo double wheel W 2k is a 3-connected planar quadrangulation with 2k + 2 vertices, while the X-pseudo double wheel XW 2k is the corresponding optimal 1-planar graph. See Figure 2(e) for an illustration of the pseudo double wheel when k = 4, and Figure 2(f) for an illustration of the X-pseudo double wheel when k = 4. Brinkmann et al. [32] proved that W 2k (k 3) is the unique series of irreducible 3-connected planar quadrangulations, and that every 3-connected planar quadrangulation can be obtained from the graph W 2k (k 3) by a sequence of vertex-splitting and 4-cycle additions. The four operations defined above for 3-connected planar quadrangulations can be easily extended to optimal 1-planar graphs. See also Figures 2(c) and 2(d) for an illustration. In particular, Suzuki [120] observed that every optimal 1-planar graph can be obtained from the graph W 2k (k 3) by a sequence of vertex-splitting and 4-cycle additions. This characterization has been used by Suzuki [121] to investigate the re-embeddability of optimal 1-planar graphs. [32] G. Brinkmann, S. Greenberg, C. S. Greenhill, B. D. McKay, R. Thomas, and P. Wollan. Generation of simple quadrangulations of the sphere. Discrete Math., 305(1 3):33 54, [36] G. J. Chang. Algorithmic Aspects of Domination in Graphs, pages Springer, New York, NY, [42] V. Chepoi, F. Dragan, and Y. Vaxès. Center and diameter problems in plane triangulations and quadrangulations. In ACM-SIAM SODA 2002, pages SIAM, [120] Y. Suzuki. Optimal 1-planar graphs which triangulate other surfaces. Discrete Math., 310(1):6 11, [121] Y. Suzuki. Re-embeddings of maximum 1-planar graphs. SIAM J. Discrete Math., 24(4): , Triangulated 1-planar graphs A map M is a partition of the sphere into finitely many regions. Each region is a closed disk and the interiors of two regions are disjoint. Some regions are labeled as countries and the nonlabeled regions are called holes. Two countries are adjacent if they touch. Given a graph G, a map M realizes G if the vertices of G are in bijection with the countries of M, and two countries of M are adjacent if and only there is an edge between the two corresponding vertices of G. A graph that can be realized as a map is called a map graph. Note that h > 1 countries meeting at a point implies the existence of a complete graph K h as a subgraph of G. If no more than h regions meet at a point, then M is an h-map and G an h-map graph. Also, if a map M has no holes, then it is called a hole-free map and its map graph G is a hole-free map graph. We refer the reader to papers by Chen et al. [37, 38, 39] and by Thorup [127] for more details and results on this family of graphs. In particular, Chen et al. [39] observed that 3-connected hole-free 4-map graphs are exactly the triangulated 1-planar graphs, that is, a 3-connected graph is triangulated 1-planar if and only if it can be realized as a hole-free 4-map. Note that the family of 1-planar graphs includes the family of 4-map graphs, but the opposite is not true [38]. The relationship between 1-planar graphs and map graphs has been further investigated by Brandenburg [26]. 9

10 [26] F. J. Brandenburg. On 4-map graphs and 1-planar graphs and their recognition problem. CoRR, abs/ , [37] Z. Chen, M. Grigni, and C. H. Papadimitriou. Planar map graphs. In STOC 1998, pages ACM, [38] Z. Chen, M. Grigni, and C. H. Papadimitriou. Map graphs. J. ACM, 49(2): , [39] Z. Chen, M. Grigni, and C. H. Papadimitriou. Recognizing hole-free 4-map graphs in cubic time. Algorithmica, 45(2): , [127] M. Thorup. Map graphs in polynomial time. In FOCS 1998, pages IEEE, Complete k-partite 1-planar graphs While it is known that the complete graph K n is 1-planar if and only if n 6, Czap and Hudák [45] characterized the complete k-partite 1-planar graphs. Specifically, K n1,n 2 is not 1- planar for n 1, n 2 5, K n1,3 is 1-planar if and only if n 1 6, and K n1,4 is 1-planar if and only if n 1 4 (recall that K n1,1 and K n1,2 are planar for all values of n 1 ). For h 7, the graph K n1,...,n h contains K 7 as a subgraph, hence it cannot be 1-planar. For 3 h 6, the characterization is based on case analysis and we let the reader refer to [45] for details. [45] J. Czap and D. Hudák. 1-planarity of complete multipartite graphs. Discrete Appl. Math., 160(4-5): , Recognition In what follows we refer to the problem of deciding whether a graph G is 1-planar as the 1- planarity problem. In general this problem in NP-hard (Section 3.2.1), although it can be solved in polynomial time in some restricted cases (Section 3.2.2). 1-planarity problem has been studied also in the setting where the input graph has a fixed rotation system (Section 3.2.3), and in terms of parameterized complexity (Section 3.2.4) NP-hardness results The 1-planarity problem is NP-complete [78, 96]. The first proof was by Grigoriev and Boadlander [78], and was based on a reduction from the 3-partition problem. We recall that the 3-partition problem is an NP-complete problem defined as follows (see also [75]). Given a set A of 3m elements, an integer B N, and an integer s(a) N for each a A such that B/4 < s(a) < B/2 and a A s(a) = mb, the problem asks if A can be partitioned into m subsets A 1,..., A m such that, for 1 i m, a A i s(a) = B. On a high level, the reduction works by constructing a graph G that separates the plane into m large regions R 1,..., R m, such that the boundaries between each region cannot be crossed in a 1-planar drawing of G. Each region R i has 3 buckets to place the elements of A i, and B buckets to place the elements of A. A second proof was later given by Korzhik and Mohar [96], who used a reduction from the 3-coloring problem for planar graphs (see [75])). [75] M. R. Garey and D. S. Johnson. Computers and Intractability: A Guide to the Theory of NP-Completeness. W. H. Freeman,

11 [78] A. Grigoriev and H. L. Bodlaender. Algorithms for graphs embeddable with few crossings per edge. Algorithmica, 49(1):1 11, [96] V. P. Korzhik and B. Mohar. Minimal obstructions for 1-immersions and hardness of 1-planarity testing. J. Graph Theory, 72(1):30 71, The 1-planarity problem remains NP-complete even for graphs with bounded bandwidth, pathwidth, or treewidth [12], and for graphs obtained from planar graphs by adding a single edge [34]. [12] M. J. Bannister, S. Cabello, and D. Eppstein. Parameterized complexity of 1-planarity. In WADS 2013, volume 8037 of LNCS, pages Springer, [34] S. Cabello and B. Mohar. Adding one edge to planar graphs makes crossing number and 1-planarity hard. SIAM J. Comput., 42(5): , Brandenburg et al. [29] proved that the problem of deciding whether a graph G is IC-planar is also NP-complete, by using a reduction from the 1-planarity problem. A similar reduction holds for NIC-planar graphs as well [10]. [10] C. Bachmaier, F. J. Brandenburg, K. Hanauer, D. Neuwirth, and J. Reislhuber. NICplanar graphs. CoRR, abs/ , [29] F. J. Brandenburg, W. Didimo, W. S. Evans, P. Kindermann, G. Liotta, and F. Montecchiani. Recognizing and drawing IC-planar graphs. Theor. Comput. Sci., 636:1 16, Polynomial-time results Brandenburg [28] recently showed that the characterization for optimal 1-planar graphs given by Suzuki [120] (see Section 3.1) can be exploited to efficiently recognize optimal 1-planar graphs. Specifically, given an n-vertex graph with 4n 8 edges, there exists an O(n)-time algorithm to decide whether G is optimal 1-planar and if G is optimal 1-planar, then the algorithm returns a valid embedding of G. [28] F. J. Brandenburg. Recognizing optimal 1-planar graphs in linear time. Algorithmica, Triangulated 1-planar graphs can be recognized in polynomial time, since these graphs correspond to 3-connected hole-free 4-map graphs (see Section 3.1), and n-vertex hole-free 4-map graphs can be recognized in O(n 3 ) time [39]. See also [26]. Note that, since optimal 1-planar graphs are triangulated, this implies that a cubic time recognition algorithm exists also for these graphs. [26] F. J. Brandenburg. On 4-map graphs and 1-planar graphs and their recognition problem. CoRR, abs/ , [39] Z. Chen, M. Grigni, and C. H. Papadimitriou. Recognizing hole-free 4-map graphs in cubic time. Algorithmica, 45(2): ,

12 Let T = (V, E T ) be a maximal plane graph with n vertices and let M = (V, E M ) be a matching. Brandenburg et al. [29] proved that there exists an O(n 3 )-time algorithm to test if G = (V, E T E M ) admits an IC-planar drawing that preserves the embedding of T. If the test is positive, the algorithm computes a feasible drawing. The interest in this special case is motivated by the fact that every IC-planar graph with maximum number of edges is the union of a triangulated planar graph and of a set of edges that form a matching [140]. [29] F. J. Brandenburg, W. Didimo, W. S. Evans, P. Kindermann, G. Liotta, and F. Montecchiani. Recognizing and drawing IC-planar graphs. Theor. Comput. Sci., 636:1 16, [140] X. Zhang and G. Liu. The structure of plane graphs with independent crossings and its applications to coloring problems. Open Math., 11(2): , Auer et al. [8] and Hong et al. [82] independently proved that recognizing outer 1-planar graphs can be done efficiently. Both proofs are based on O(n)-time algorithms that decides whether an n-vertex input graph G admits an outer 1-planar embedding. In the positive case both algorithms return a valid embedding of G. In the negative case, the algorithm described in [8] also returns one of six possible minors that are not outer 1-planar. [8] C. Auer, C. Bachmaier, F. J. Brandenburg, A. Gleißner, K. Hanauer, D. Neuwirth, and J. Reislhuber. Outer 1-planar graphs. Algorithmica, 74(4): , [82] S. Hong, P. Eades, N. Katoh, G. Liotta, P. Schweitzer, and Y. Suzuki. A linear-time algorithm for testing outer-1-planarity. Algorithmica, 72(4): , Triangulated IC-planar and NIC-planar graphs can be recognized in cubic time [27], while optimal NIC-planar graphs can be recognized in linear time [10]. [10] C. Bachmaier, F. J. Brandenburg, K. Hanauer, D. Neuwirth, and J. Reislhuber. NICplanar graphs. CoRR, abs/ , [27] F. J. Brandenburg. Recognizing IC-planar and NIC-planar graphs. CoRR, abs/ , Fixed rotation system setting The recognition problem has also been studied with the additional assumption that the input graph comes along with a rotation system. Deciding whether a graph G with a given rotation system R admits a 1-planar drawing that respects R is NP-complete, even if G is 3-connected [9]. On the positive side, Eades et al. [66] proved that there is an O(n)-time algorithm to decide if an n-vertex graph G with a rotation system R has a maximal 1-planar embedding (i.e., a 1-planar embedding in which no edge can be added without violating 1-planarity) that respects R. In the positive case, the algorithm returns a valid embedding of G. The algorithm is based on the following two properties (proved in [66]). First, in any maximal 1-planar embedding, the subgraph induced by the edges that do not intersect any other edge is spanning and 2-connected. Second, if G admits a maximal 1-planar embedding that respects R, then the embedding is unique. [9] C. Auer, F. J. Brandenburg, A. Gleißner, and J. Reislhuber. 1-planarity of graphs with a rotation system. J. Graph Algorithms Appl., 19(1):67 86, [66] P. Eades, S.-H. Hong, N. Katoh, G. Liotta, P. Schweitzer, and Y. Suzuki. Testing maximal 1-planarity of graphs with a rotation system in linear time. In GD 2012, volume 7704 of LNCS, pages Springer,

13 3.2.4 Parameterized complexity Given its NP-hardness, it is natural to study the 1-planarity problem in terms of parameterized complexity. That is, we seek additional parameters (other than the numbers of edges and vertices) that measure the complexity of an input graph, and design recognition algorithms whose running time is the product of a polynomial in the input size and a non-polynomial function of these additional parameters. For more details on parameterized complexity theory, see [61, 73]. In this direction, Bannister et al. [12] proved that, for an n-vertex graph G, deciding whether G is 1-planar is fixed parameter tractable for various graph parameters. More precisely, the problem can be solved in time O(n + 2 O(k2) ) if G has cover number k; in time O(n2 22d2 +O(d)) if G has tree-depth d; and in time O(n + 2 O((3c)!) ) if G has cyclomatic number c. [12] M. J. Bannister, S. Cabello, and D. Eppstein. Parameterized complexity of 1-planarity. In WADS 2013, volume 8037 of LNCS, pages Springer, [61] R. G. Downey and M. R. Fellows. Parameterized Complexity. Monographs in Computer Science. Springer, [73] J. Flum and M. Grohe. Parameterized Complexity Theory. Texts in Theoretical Computer Science. An EATCS Series. Springer, Structural properties In this section we deal with structural properties and invariants of 1-planar graphs. In Section 4.1, we review results on the classic coloring problem. In Section 4.2, we report bounds on the number of edges of 1-planar graphs. Section 4.3 contains recent results on the problem of partitioning the edges of a 1-planar graph, such that each partition set induces a graph with predefined properties. In Section 4.4 we describe bounds on various graph parameters of 1-planar graphs. Section 4.5 contains results about the existence of subgraphs of bounded vertex degree in 1-planar graphs. Finally, results related with binary graph operations that preserve 1-planarity are covered in Section Coloring A coloring of a graph G is an assignment of labels called colors to elements of G subject to certain constraints. Based on whether the elements of G to be colored are its vertices, its edges, or both, we distinguish among vertex coloring, edge coloring, and total coloring, respectively. Coloring problems find application in task scheduling and frequency assignment, register allocation, as well as in pattern matching, designing seating plans, solving Sudoku puzzles and others; see for example: [35] G. J. Chaitin. Register allocation & spilling via graph coloring. SIGPLAN Not., 17(6):98 101, [99] R. Lewis. A Guide to Graph Colouring: Algorithms and Applications. Springer, [103] D. Marx. Graph colouring problems and their applications in scheduling. Periodica Polytechnica Ser. El. Eng., 48(1):11 16,

14 4.1.1 Vertex coloring A proper vertex coloring of a graph G is an assignment of colors to the vertices of G such that no two adjacent vertices receive the same color. The smallest number of colors needed to obtain a proper vertex coloring of a graph G is called the chromatic number of G. A classic result in graph theory is that every planar graph has chromatic number at most 4: [7] K. Appel and W. Haken. Every planar map is four colorable, volume 98 of Contemporary Mathematics. AMS, Ringel [115] proved that every 1-planar graph has chromatic number at most 7, and conjectured that this bound could be lowered to 6. Borodin [23] settled the conjecture in the affirmative, proving that the chromatic number of each 1-planar graph is at most 6. The bound is tight as for example the complete graph K 6 is 1-planar and requires six colors. The same author later showed a relatively shorter proof of this result [24]. However, Borodin s proof does not lead to an efficient algorithm for computing a proper vertex coloring with 6 colors of a 1-planar graph. On the other hand, there is an O(n)-time algorithm for computing a proper vertex coloring with 7 colors of any n-vertex 1-planar graph [40]. If we restrict to IC-planar graphs, then Král and Stacho [97] proved that the chromatic number of IC-planar graphs is 5. Finally, it is NP-complete to decide whether a given 1-planar graph admits a proper vertex coloring using four colors [40]. [23] O. V. Borodin. Solution of the Ringel problem on vertex-face coloring of planar graphs and coloring of 1-planar graphs. Metody Diskret. Analiz, 108:12 26, [24] O. V. Borodin. A new proof of the 6 color theorem. J. Graph Theory, 19(4): , [40] Z. Chen and M. Kouno. A linear-time algorithm for 7-coloring 1-plane graphs. Algorithmica, 43(3): , [97] D. Král and L. Stacho. Coloring plane graphs with independent crossings. J. Graph Theory, 64(3): , [115] G. Ringel. Ein Sechsfarbenproblem auf der kugel. Abhandlungen aus dem Mathematischen Seminar der Universitaet Hamburg, 29(1 2): , Other forms of vertex coloring have been studied. An acyclic coloring of a graph G is a proper vertex coloring of G such that every cycle of G uses at least three colors. The smallest number of colors needed to obtain an acyclic proper vertex coloring of a graph G is called the acyclic chromatic number of G. Borodin et al. [22] proved that every 1-planar graph has acyclic chromatic number at most 20 and there is a 1-planar graph which requires at least 7 colors in any acyclic coloring. [22] O. Borodin, A. Kostochka, A. Raspaud, and E. Sopena. Acyclic colouring of 1-planar graphs. Discrete Appl. Math., 114(13):29 41, Given a graph G and a set L(v) of colors for each vertex v of G (called a list), a list coloring is a proper vertex coloring of G such that each vertex is assigned with a color in its corresponding list. Wang and Lih [129] proved that 1-planar graphs have list colorings with at most seven colors. [129] W. Wang and K.-W. Lih. Coupled choosability of plane graphs. J. Graph Theory, 58(1):27 44,

15 4.1.2 Edge coloring A proper edge coloring of a graph G is an assignment of colors to the edges of G such that no two adjacent edges receive the same color. The smallest number of colors needed to obtain a proper edge coloring of a graph G is called the chromatic index of G. By Vizing s theorem [128], the number of colors needed to edge color a graph G with maximum vertex degree is either or + 1. [128] V. G. Vizing. On an estimate of the chromatic class of a p-graph. Diskret. Analiz No., 3:25 30, Zhang and Wu [143] proved that every 1-planar graph with maximum vertex degree 10 has chromatic index. Also, every 1-planar graph without adjacent triangles and with maximum vertex degree 8 has chromatic index [138]. Moreover, every 1-planar graph without chordal 5-cycles and with maximum vertex degree 9 has chromatic index, and for each 7 there exist 1-planar graphs with maximum vertex degree and chromatic index + 1 [139]. The chromatic index of an IC-planar graph G with maximum vertex degree 8 is [140]. The chromatic index of outer-1-planar graphs with maximum vertex degree 4 is, and there are infinitely many outer 1-planar graphs with maximum vertex degree 3 and chromatic index 4 [141]. Based on these results, Zhang [136] argues that the chromatic index of any outer 1-planar graph can be determined in polynomial time. [136] X. Zhang. The edge chromatic number of outer-1-planar graphs. CoRR, abs/ , [138] X. Zhang and G. Liu. On edge colorings of 1-planar graphs without adjacent triangles. Inf. Process. Lett., 112(4): , [139] X. Zhang and G. Liu. On edge colorings of 1-planar graphs without chordal 5-cycles. Ars Comb., 104: , [140] X. Zhang and G. Liu. The structure of plane graphs with independent crossings and its applications to coloring problems. Open Math., 11(2): , [141] X. Zhang, G. Liu, and J.-L. Wu. Edge covering pseudo-outerplanar graphs with forests. Discrete Math., 312(18): , [143] X. Zhang and J.-L. Wu. On edge colorings of 1-planar graphs. Inf. Process. Lett., 111(3): , The list variant of the proper edge coloring problem has also been considered. Specifically, given a graph G and a set L(e) of colors for each edge e of G (called a list), a list edge coloring is a proper edge coloring of G such that each edge is assigned with a color in its corresponding list. Every 1-planar graph with maximum vertex degree has a list edge coloring with + 1 colors if 16 and with colors if 21 [142]. If G is IC-planar and 8, then it has a list edge coloring with colors [140]. [140] X. Zhang and G. Liu. The structure of plane graphs with independent crossings and its applications to coloring problems. Open Math., 11(2): , [142] X. Zhang, J. Wu, and G. Liu. List edge and list total coloring of 1-planar graphs. Front. Math. China, 7(5): ,

16 4.1.3 Total coloring A total coloring of a graph G is an assignment of colors to the vertices and edges of G such that neither two adjacent edges, two adjacent vertices nor an edge and its end-vertices receive the same color. The smallest number of colors needed to obtain a total coloring of a graph G is called the total chromatic number of G. Behzad [14] conjectured that the total chromatic number of a graph G with maximum vertex degree is at most + 2. [14] M. Behzad. Graphs and Their Chromatic Numbers. PhD thesis, Michigan State University, Department of Mathematics, Zhang et al. [137] proved that the total coloring conjecture holds for 1-planar graphs with maximum vertex degree 13. Also, Czap [44] proved that the total chromatic number of a 1-planar graph G with maximum vertex degree 10 is at most + 3. [44] J. Czap. A note on total colorings of 1-planar graphs. Inf. Process. Lett., 113(14-16): , [137] X. Zhang, J. Hou, and G. Liu. On total colorings of 1-planar graphs. J. Comb. Optim., 30(1): , The list variant of the total coloring problem has also been studied. Specifically, given a graph G and a set L(e) (respectively, L(v)) of colors for each edge e of G (respectively, vertex v of G), a list total coloring is a total coloring of G such that each edge and each vertex is assigned with a color in its corresponding list. Every 1-planar graph with maximum vertex degree has a list total coloring with + 2 colors if 16 and with + 1 colors if 21 [142]. If G is IC-planar and 8, then it has a list total coloring with + 1 colors [140]. [140] X. Zhang and G. Liu. The structure of plane graphs with independent crossings and its applications to coloring problems. Open Math., 11(2): , [142] X. Zhang, J. Wu, and G. Liu. List edge and list total coloring of 1-planar graphs. Front. Math. China, 7(5): , Edge density The fact that a 1-planar graph G with n vertices has at most 4n 8 edges has been proved multiple times [19, 41, 72, 111]. This bound is shown to be tight for every n = 8 and for n 10 [19]. Given the characterization of optimal 1-planar graphs discussed in Section 3.1, this result can be easily observed as follows. Let G be an optimal 1-planar graph with n vertices, and consider the underlying 3-connected planar quadrangulation Q of G. Graph Q has n 2 faces and 2n 4 edges. Inserting a pair of crossing edges in each face of Q adds 2(n 2) edges, which leads to 4n 8 edges in total. An optimal 1-planar graph is shown in Figure 1(b). [19] R. Bodendiek, H. Schumacher, and K. Wagner. Bemerkungen zu einem Sechsfarbenproblem von G. Ringel. Abhandlungen aus dem Mathematischen Seminar der Universitaet Hamburg, 53(1):41 52, [41] Z.-Z. Chen. New bounds on the edge number of a k-map graph. J. Graph Theory, 55(4): , [72] I. Fabrici and T. Madaras. The structure of 1-planar graphs. Discrete Math., 307(7 8): ,

17 (a) (b) (c) Figure 3: (a) A maximal 1-planar graph with 20 vertices and 44 edges. (b) A maximal IC-planar graph with 8 vertices and 20 edges. (c) A maximal NIC-planar graph with 12 vertices and 36 edges. [111] J. Pach and G. Tóth. Graphs drawn with few crossings per edge. Combinatorica, 17(3): , If we consider only those 1-planar graphs that admit a 1-planar straight-line drawing (see Section 5), then any such graph with n-vertices has at most 4n 9 edges, as proved by Didimo [57] (see also the alternative proof by Ackerman [1]). [1] E. Ackerman. A note on 1-planar graphs. Discrete Appl. Math., 175: , [57] W. Didimo. Density of straight-line 1-planar graph drawings. Inf. Process. Lett., 113(7): , Density considerations of 1-planar graphs are fundamentally different from the equivalent considerations for planar graphs. For example, Brandenburg et al. [30] proved that there exist maximal 1-planar graphs with n vertices and n 17 < 2.65n edges; see Figure 3(a) for a sparse maximal graph obtained with the construction in [30]. This is in contrast with maximal planar graphs, which have exactly 3n 6 edges. Moreover, any n-vertex maximal 1-planar graph has at least n 10 3 edges [30]. Brandenburg et al. [30] also studied the problem in the fixed rotation system setting. They proved that, for infinitely many values of n, there are n-vertex maximal 1-planar graphs with a fixed rotation system that have 7 3n 3 edges, and that any maximal 1-planar graph with a fixed rotation system has at least n 10 3 edges. [30] F.-J. Brandenburg, D. Eppstein, A. Gleißner, M. T. Goodrich, K. Hanauer, and J. Reislhuber. On the density of maximal 1-planar graphs. In GD 2012, volume 7704 of LNCS, pages Springer, An IC-planar graph G with n vertices has at most 3.25n 6 edges and this bound is tight [140]. The graph in Figure 3(b) is used in [140] to prove the tightness of the result. A NIC-planar graph G with n vertices has at most 3.6n 7.2 edges and this bound is tight [49] (see also [10]). For example, Figure 3(c) shows a maximal NIC-planar graph that matches this bound. [10] C. Bachmaier, F. J. Brandenburg, K. Hanauer, D. Neuwirth, and J. Reislhuber. NICplanar graphs. CoRR, abs/ ,